中文摘要：

分析稀少数据的相关性是一种重要的、有价值的数据挖掘任务。运用面向关联规则的FP树构造方法，提出了一种特异关联规则挖掘算法RSFPA。该算法将包含特异模式的数据集压缩成一棵FP树，通过挖掘FP树来提取特异模式集，从而进一步提高了特异模式的挖掘效率。最后，利用恒星光谱作为数据集，实验验证了RSFPA算法的正确性和有效性。

英文摘要：

Interrelation analysis of rare data is an important and valuable task in data mining. Peculiarity association rule mining algorithm RSFPA based on FP-tree is presented by using FP-tree idea oriented association rule miming. The Algorithm compresses date set contained peculiarity pattern into a FP-tree , and mines peculiarity pattern from the FP-tree ,therefore to improve mining efficiency of peculiarity pattern. In the end, the RSFPA algorithm is validated by using star spectrum data as experiment data.